Dr. Sarath GuttikundaUrbanEmissions.info

A Primer on Source Apportionment (2011)

Author: Dr. Sarath GuttikundaDesigned by: Puja Jawahar

DISCLAIMER:All characters are fictional.

Errors & interpretations are the sole responsibility of the authors.

(c) www.urbanemissions.info

Source Apportionment .................................................1

Top-Down Approach .....................................................2Ambient Sampling .............................................3Laboratory Analysis ........................................ .4Receptor Modeling ............................................5

Bottom-Up Approach ................................................. 8Emissions Inventory ........................................ 9Dispersion Modeling ........................................12

Reconciling Results .........................................................14

In Summary ......................................................................15

Further References .......................................................16

How do we reduce pollution in the city?

1

I already have an idea of the sources of air pollution

here. Why do i need a study?

Are the results of the 2 methods similar?

first we need to conduct a source apportionment study which will tells us what the sources of pollution are and

how much each source contributes to total pollution.

because a study will scientifically determine these contributions. Thus eliminating the tendency to

overemphasize certain sources or underemphasize or ignore others.

there are 2 ways to conduct source apportionment.

1) a Top-Down Approach, which involves collecting samples and analyzing them in a lab.

and/or2) A bottom-Up ApproAch, which uses existing data and

survey methods.

ideally yes.

Using the Top-down approach, we get accurate results for specific locations (where samples are collected),

which we then average to get a city-level profile.

Using the Bottom-up approach, depending on the existing data, we get pollution estimates for any part

of the city.

City Representativesim-air man (facilitator)

2

(1)Decision to

conduct source apportionment

(2)Ambient

Sampling

(3)Laboratory

Analysis

(4)Receptor modeling

this approach measures pollution using filters. these are then

analyzed in the lab for chemical signatures to assess source

contributions.

The steps seem

straight forward, but I am sure each one is

a project in itself...

This method is applied for particulate pollution only. Ambient sampling is carried out for two size fractions - PM10 (coarse PM) and Pm2.5 (Fine PM)

how many samples do we collect?

The more the merrier... Keeping budget constraints in mind, an ideal study would include a mix of residential,

industrial, roadside and background locations and

collect at least 15 samples from each location every

season.

3

why do we need to conduct sampling for

every season?

while some of the sources remain the same every season, fuel consumed and resulting

pollution varies ...

meteorological conditions also vary

between seasons, which affects the ambient

pollution.

A multi-season study thus provides more accurate

results on average.

For a particular season, sampling should be conducted simultaneously at all locations.

This will allow for comparison of pollution levels and sources between locations.

for example, domestic and

industrial heating requirements are different between

summer and winter.

Can you tell us about a commonly used

sampler?

we need filter based samplers

(see the picture for an example of a mini-vol sampler - these are

portable, easy to handle and are Battery operated)

An example of a sample

filter for storage

4

once collected, what do we do with the

filter samples?

WHY DO WE NEED SO MUCH LAB WORK?

THE FILTERS are STORED IN A CONTROLLED ENVIRONMENT (to PROTECT THE SAMPLES) and sent to the laboratory for analysis.

WE NEED INFORMATION ON METALS, IONS, AND CARBON to

successfully determine the source contributions - THERE IS

NO SINGLE LABORATORY EQUIPMENT TO ANALYZE THEM ALL.

THE lab equipment is NOT CHEAP. HOWEVER, ONCE THE SYSTEMS ARE IN PLACE AND WITH PROPER TRAINING, ONE COULD REPLICATE

THESE STUDIES.

IF a city does not operate its OWN LABS, they can send the samples to a CERTIFIED LAB.

HOW EXPENSIVE is this ANALYsis?

5

The most common receptor model used for this kind of analysis is the

Chemical Mass Balance (cmb) model.

besides the lab analysis, we need source profiles.

Expert on modeling what are source

profiles?

A source profile looks similar to the results of an ambient

sample, but is specific to the source type

It is developed by collecting

samples very close to the source, in order to capture

the source's "signature"

Ideally, one should have source profiles from the city

where the study is being conducted. However, if budgets are limited, we can borrow source profiles from other cities with similar pollution

sources.

common markers include

Al & Si for road dustK for Biomass

Na for sea-saltA mix of metals for coal

BC & oc ratios for vehicle exhaust

6

An example Source

Profile ...

These sound like challenging (and

expensive!) steps...

This approach provides

pollution information specific to the sampling

location.

For example, a sample collected near a traffic

junction will identify transportation as the largest source of

pollution.

Each filter sample only reflects the sources that are within a 2-3 km radius of the sampling site. Therefore, by averaging the results from

several locations you get an approximation of city-level profile.

7

no. This method assesses source apportionment for

particulate pollution only.

can we repeat this experiment for all

pollutants?

in Summary, when all the protocols are followed, this top-down approach is scientifically robust and provides

credible results.

At the same time, this can get pretty expensive, provides limited information (depending on how many samples we collect), and is

applicable for particulate pollution only.

But the city is expanding fast and resources are limited. plus there might be non-particulate pollutants such as So2, NOx, Co, and

Ozone, which are also at critical levels.

In this case, what are our alternatives?

That is a good observation.

In addition to following all protocols carefully, one also has to ensure that the

analysis is conducted by a reputable laboratory.

There is an alternative approach that I mentioned earlier, which uses existing data and survey

methods.

We will discuss this approach in the next section - Bottom-up approach

NOtes and Questions...

8

(1)Decision to

conduct modeling

(2)emissions inventory

(3)dispersion modeling

A bottom-up approach estimates pollution based on the distribution of pollution

causing activities in a city

it is really important that every city establishes an

emissions inventory.

This serves as a baseline for many pollution and urban

policy related action plans.

a Basic inventory can be developed using

existing information and minimum resources.

First, you map various pollution-causing

activities*. Then you zero in on each of these sources to

estimate energy consumption and emissions.

* Commonly addressed sources are

industry - cement, bricks, power plants, etc

residential cooking and heating

garbage burning

vehicle exhaust and road dust

construction and

small diesel power generator sets

9

** Emission factor: Unit of pollution released per unit of energy consumed. Thus... emissions = emission factor * Energy

compile industrial energy consumption

through audits & energy statistics

.... Analyze vehicle statistics from

transport census for Vehicle miles traveled, no of vehicles, etc...

We can then use existing Emission

factors to calculate emissions

by sector

.... Survey the domestic sector to get a sense of energy use at the household level...

what kind of activity data is required to create an emissions

inventory?

what other sources should we consider?

Seasonal sources such as dust storms, forest fires, and agricultural burning also contribute to pollution.

Once we have gathered this data about inputs/energy use by sector, we can use

existing emission factors**...

10

Not really.

emission factors do vary depending on the combustion

technologies in use in the city - but Usually fall within a

range.

If you have budget and time constraints, using existing

numbers is a good way to start.

This can always be improved upon time

and resource permitting.

We do not have access to city specific emission factors. DOes it affect results if we use existing

emission factors from the Literature versus precise

factors?

Usually INventories are established for particulates (pm), Sulfur dioxide (So2), nitrogen oxides (nox), carbon monoxide (co), hydrocarbons (voc's), black carbon (BC), and organic carbon (OC) - tons/year

THATS RIGHT !

AT THE INDUSTRY LEVEL, A LOT OF SPECIFIC DETAILS

ARE REQUIRED.

BUT, AT THE URBAN AND REGIONAL LEVELS,

AGGREGATE INPUTS SUFFICE.

11

our emissions inventory for particulates give us the percentage

contribution of various sources.. and we also have the Percentage

contribution from the top-down approach.

What is the difference between the two?

that's a very important question.

those two results are in fact very different.

Results of an emissions inventory give the weight of

pollution from various sources (mass/year)

results of a top-down study give the source contributions to the ambient concentrations

(mass/volume).

Using A dispersion model (and local meteorological

conditions), we can convert the emissions into ambient

concentrations.

These results can then be compared to the results of the

top-down approach.

the Type of sources influence ambient concentrations.

For example, ground level emissions from vehicle exhaust, though a small % in the

inventory have a disproportionately larger share in the local concentrations.

while, a power plant emitting a lot of pollution, contributes less to the immediate vicinity due to

long range transport.

that makes

sense..

% Emissions does not equal % concentrations

12

once we have our emissions inventory,

dispersion modeling allows us to get a spatial picture

of pollutant "concentrations"

Dispersion modeling is a technically

challenging step, which requires a certain level of

training and proficiency. Often this step becomes a bottle-neck

for cities trying to map pollution.

see the comparison of models below.

An illustration of a dispersion modeling output ....

The output of this step is most valuable, since the gridded

concentrations will provide us with a pollution map, along with hot spot

locations, for the entire city.

For example,

models like GAINS, HEAT, SEI toolkit, and USEPA's AP-42, can help with emissions

inventory;

models like CMAQ, WRF, ATMOS, and ISC3, can help with dispersion modeling.

After careful consideration of requirements and resources, one should

choose an appropriate model.

Are there any specific models which we can

follow?

13

should we model all pollutants?

but, the chemistry of the pollutants is interlinked.

Can I use this analysis for health impact

assessments?

not necessarily.

You can focus on the critical pollutants, according to the purpose of the pollution analysis. for example, for public health concerns fine pM is critical, for acid rain concerns - so2 and nox,

visibility - smog producing ozone, and so on..

But if you have the resources, institutional capacity, and data, you can model all pollutants and get a comprehensive assessment of

pollution in the city.

That is correct..

chemistry of pollutants is interlinked, which can be included in the multi-pollutant

dispersion modeling.

certainly.

The gridded concentrations, along with gridded population over the city, can be used for

mortality and morbidity calculations.

A major concern is in the emissions inventory step. we need a good

understanding of the geographical spread of the major sources of pollution.

14

Note that each approach can be used

to strengthen the other.

for example, while building source profiles, measurements conducted

close to the sources can validate the emissions

factors, which can be used to build a robust

emissions inventory

Results based on the top-down approach can be used to correct the missing sources in the emissions inventory.

The dispersion modeling results can help identify

the pollution hot spots in the city, where sampling

for the top-down approach can be performed.

monitoring data can help validate the results of

dispersion modeling, which can be further expanded for evaluating "what-if"

emission scenarios.

Together, these methods provide a strong resource for developing a pollution

control strategy.

15

Excellent.

These are two distinct approaches to come to the

same conclusions.

Depending on our technical and financial resources, we can utilize either approach

(or both).

indeed.. thank you !!

Exactly.

Both the methods are robust and well tested in cities

across the world.

THe most important step is taking the decision to conduct a

Source apportionment study. Then, based on your budget and the information that is already available, we can decide which approach is most appropriate

For more details and case studies, you can visit us@ http://www.urbanemissions.info

or

send me an emailsimair@urbanemissions.info

16

A handbook published by ESMAP (the World Bank, 2011) on the techniques of source apportionment, along with an array of applications from across the world is available @ http://www.esmap.org/esmap/node/1159

An application of top-down and bottom-up approaches for the city of Hyderabad, India, was conducted under the US-

EPA's IES program in 2007-08, is available@ http://www.epa.gov/ies/india/apportionment.htm

An application of the bottom-up approach in Delhi, India, where an emissions inventory was built, as part of an air quality forecasting system is available @ http://www.aria.fr/delhi

The "Chemical Mass Balance" (CMB) is the most common receptor model in use for a number of source apportionment

studies across the world. An overview of the model, along with the merits and limitations are explained in detail

@ http://www.epa.gov/scram001/receptor_cmb.htm

To estimate emissions inventory, conduct dispersion modeling, and assess health impacts, access the the SIM-air family of tools @ http://www.urbanemissions.info

www.urbanemissions.info